Refining neutropenia risk assessment in patients treated with first-line endocrine therapy (ET) and cyclin-dependent kinase 4/6 inhibitors (CDK4/6i) for metastatic breast cancer (MBC) through a cell-free DNA workflow (cfDNA)

1027

Background: The combination of ET and CDK4/6i is the current standard of care for hormone receptor (HR)-positive/HER2-negative MBC (luminal MBC), with neutropenia being the main dose limiting toxicity. We previously observed the potential association between leucocyte count (WBC) and different fractions of cell free circulating DNA (cfDNA) (Bortot et al 2020). The present study aimed to evaluate the feasibility of a cfDNA-based workflow as a new tool to assess the risk of treatment induced neutropenia. Methods: The study analyzed a prospective cohort 83 luminal MBC patients (pts) treated with first line ET and CDK4/6i in the CRO-2018-56 multicenter study from 2018 to 2021. cfDNA was characterized through droplet digital PCR (ddPCR) based on different ACTB DNA fragments lengths: short (s), medium (m) and long (l). Blood samples were collected before treatment start (BL) and at the first clinical evaluation after 3 months (E1). Associations between clinical characteristics, neutropenia and cfDNA were explored through Kruskal Wallis, time to G3 neutropenia (NG3) (TTN) was analyzed through log-rank and Cox regression. Results: Neutropenia was G3 in 44 out of 83 pts (53%) and G4 in 2 pts (2%). Median TTN was 1.8 months, 60% of NG3 occurred within 3.7 months. Overall, 74 pts (89%) resolved toxicity within 7 days, 10 pts (12%) reduced CDK4/6i dose after NG3. BL neutrophils count (Neu) and WBC were significantly lower in pts that developed NG3 (P = 0.0013 and P = 0.0020 respectively). De novo metastatic pts had numerically higher Neu (median 4825 vs 3895), but only a numerically lower risk of NG3 was observed (HR 0.53 P = 0.064). Although bone involvement was not associated with risk of developing NG3, the total number of metastatic lesions was associated with higher NG3 (P = 0.0016). In particular, > 5 metastatic lesions were associated with higher NG3 risk (p = 0.013). E1 ACTB_m was significantly lower with respect to BL in pts that experienced NG3 (median 100% vs 16%, P = 0.0136 in NG3 no vs yes) with a consistent impact on the risk of NG3 (HR: 2.81, P = 0.025). No associations were observed for the other ACTB fragment length fractions. BL Neu and ACTB_m dichotomized at the median were then combined to describe 4 distinct TTN risk groups (P = 0.0006). Interestingly, pts with low BL Neu and low E1 ACTB_m had a median TTN of 0.9 months, while pts with high BL Neu and high E1 ACTB_m have not experienced NG3 after a median follow-up of 16.1 months. Conclusions: The present study proofed the concept of using cfDNA to provide clinically meaningful data not only about tumor biology, but also for a comprehensive patient assessment. Based on these results, a prospective study focused on a multiparametric neutropenia risk assessment will be started.